How to Catch Detractors Before the CSI Survey Is Submitted

What Is the OEM Survey Window — and Why Does It Define Your CSI Exposure?

OEM satisfaction surveys typically ship within 48–72 hours of the repair order closing, with most customers receiving them 3–10 days after their service visit (J.D. Power, 2026). That window is the only period where you can influence the score — once the customer clicks submit, the number goes into your CSI calculation with no appeal and no second draft. Every incentive, allocation, and certification consequence attached to that score flows from a decision you can no longer change.

Here's the part that compounds the problem. Unhappy customers don't wait politely for your follow-up call. A customer who left with a repair they don't trust, or who waited three hours for a job quoted at 90 minutes, is the most motivated survey respondent in your queue. They open the email, fill it out, and hit submit — often before a BDC rep has dialed their number. A satisfied customer may take two days to get around to the survey. A detractor is typically done within hours of receiving it.

This is why I frame the CSI challenge as a coverage problem before it's a quality problem. If your team is reaching 15–20% of customers within the first 24 hours — the realistic ceiling for a manual BDC (Lokam network data, 2025–2026) — then the detractor who received their survey email on day two has almost certainly already submitted before anyone from your store called. You didn't lose that score because of the service failure. You lost it because the recovery call didn't happen inside the window.

“The survey window is the only period where you can influence the score. Once a customer clicks submit, the number is locked — and every incentive, allocation, and certification consequence follows from a decision you can no longer change.”

How Do You Identify Detractors Before They Submit?

The signals that predict a detractor are almost always present before the follow-up call happens — they just aren't being read systematically. J.D. Power's 2026 U.S. Customer Service Index Study found that only 26% of customers experience 9 or 10 of the top 10 CSI key performance indicators. When all 10 are met, satisfaction averages 979 out of 1,000. When only 3 are met, it drops to 632 (J.D. Power, 2026). Every missed KPI on a repair order is a signal — and those signals live in your DMS before anyone picks up the phone.

There are four signals I tell dealers to flag at the repair order level, before the follow-up queue is built. First: first-visit fix failure. J.D. Power's 2025 CSI Study found that 12% of repairs aren't completed correctly on the first visit, and only half of those customers planned to return (J.D. Power, 2025). A returned RO or a same-day callback about an unresolved issue is your highest-confidence detractor signal — it's not ambiguous. Second: service time overruns. When a job quoted at two hours takes five, the customer's frustration is structural, not just emotional — it disrupted their day in a way they remember. Third: advisor notes flagging a complaint at write-up or pickup. These exist in CDK and VinSolutions for stores that capture them, but most BDCs never pull advisor notes before dialing. Fourth: repeat warranty work on the same issue. A customer returning for the same problem a second time is at acute CSI risk regardless of how the current visit went.

The difference between a manual process and an AI-assisted one is what happens during the call itself. A BDC rep can surface these signals if they're trained and briefed before each call — but at 15–20% contact rate, most flagged ROs never receive a call inside the window anyway. When AI voice outreach handles first-contact and reaches 65–75% of customers, it listens for signals in real time: complaint language, tone shifts, references to unresolved problems. A flagged call goes to a service manager within minutes. That's what catching 180 detractors looks like in practice — not a spreadsheet exercise, but a contact rate and escalation speed problem solved simultaneously.

“The signals that predict a detractor are almost always in the repair order before anyone dials. The challenge isn't identifying them — it's reading them fast enough to act inside the survey window.”

What Does an Effective Service Recovery Call Actually Look Like?

The service recovery call is not a customer satisfaction survey. That distinction matters because most BDC reps are trained to collect feedback, not to resolve a complaint in real time. When a detractor answers the phone, they aren't looking for someone to record their grievance — they're deciding whether the dealership cares enough to fix the problem. The goal of the recovery call is to move from that decision point to a resolved outcome before the OEM survey email gets opened.

Three things determine whether a recovery call works. First: who makes it. A service advisor calling to check in doesn't carry the authority to fix a problem — they're the person the customer is already frustrated with. A service manager calling to take personal ownership of an issue is an entirely different conversation. The customer's mental model of who can actually help them shifts the moment a manager is on the line. TARP's research — one of the most replicated findings in service recovery literature — shows that 54–70% of customers who lodge a complaint will return if their issue is resolved, rising to 95% when it's resolved the same day the complaint surfaces (TARP). Authority and speed are the two levers. Both require a fast escalation path from first contact to manager callback.

Second: the frame of the call. An opening like 'I just wanted to make sure everything was okay with your visit' invites a polite brush-off. An opening like 'I'm calling about the brake job on your Camry last Tuesday — I saw a note that you had a concern after pickup, and I wanted to make sure we got that resolved for you' is specific, personal, and signals that someone actually read the notes. Third — and this one matters for OEM compliance — the call should never ask a customer to delay submitting the survey or give it a higher score. That's a violation of OEM survey integrity and a credibility risk if the customer perceives it as pressure. The only job of the recovery call is to fix the problem.

“A service advisor calling to check in doesn't carry the authority to fix a problem. A service manager calling to take personal ownership does — and that distinction is what determines whether the recovery call actually works.”

What Does Lokam's Escalation Data Show About Detractor Recovery?

From Lokam's escalation data across the dealerships we work with, recoverable detractors share a consistent profile. They flagged a specific, fixable problem: a noise that returned after the repair, an advisor who didn't follow through on a commitment, a wait time that wasn't communicated. A manager callback that acknowledges the specific issue and commits to a concrete resolution — a return visit, a partial refund, a follow-up call from a named person — converts at a measurably higher rate than a generic 'sorry about your experience' call.

The detractors who are typically not recoverable within the survey window are those who experienced a first-visit fix failure on a significant repair and don't trust the vehicle. Their frustration isn't addressable by a phone call alone — it requires physically resolving the repair and demonstrating the fix, which can't happen inside a two-hour window. For these customers, the recovery call still matters for the long-term relationship, but it rarely changes the survey outcome. Being honest with service managers about this distinction is important: not every escalation ends in a converted detractor, and a process that treats all escalations identically misallocates manager time.

The broader pattern the data shows is consistent with TARP's research: when a complaint is surfaced quickly and a manager with authority responds the same day, the recovery rate is meaningful. Cox Automotive's 2025 data found that 45% of vehicle owners are dissatisfied with their dealership service experience — and that dealerships have lost 12% of service visits to competition since 2018 (Cox Automotive, 2025). The majority of those customers don't become negative surveys because of unrecoverable service failures. They become negative surveys because no one reached them in time.

“Recoverable detractors flagged a specific, fixable problem. A manager callback that acknowledges the exact issue and commits to a concrete resolution converts at a measurably higher rate than any generic follow-up call.”

How to Build a Manual Detractor Triage Process

For dealers who aren't yet using AI-assisted outreach, a manual triage process can meaningfully improve detractor coverage — provided volume stays manageable. The process starts at the repair order level, not at the follow-up queue. Before the BDC dials a single call, service managers should flag ROs in four categories: first-visit fix failures, excessive wait time (jobs more than 50% over quoted time), advisor escalations logged in CDK or VinSolutions, and repeat warranty work on the same issue. These flagged ROs move to the front of the follow-up queue regardless of how the standard call list is sorted — because they have the shortest effective window before the survey ships.

The follow-up call on a flagged RO is not a BDC job. This is the most common implementation failure: the BDC makes a courtesy call, the customer expresses frustration, the rep logs it and moves on. The triage process only works if a complaint triggers an immediate handoff to a service manager — within the same call, or within two hours. The BDC rep's job is to surface the complaint. The manager's job is to resolve it.

Build the escalation path before you run the process. Define who the designated manager is for same-day callbacks on flagged escalations, what authority that manager has — schedule a return visit, issue a partial refund, escalate to the GM — and what the closed-loop confirmation looks like in the CRM: manager called, what was offered, what the customer said. Without the closed loop, you're measuring call volume, not outcomes. The triage process is only as strong as the escalation it feeds.

“The triage process only works if a complaint triggers an immediate handoff to a service manager. A BDC rep's job is to surface the complaint. The manager's job is to resolve it.”

Why Manual Triage Fails Above 200 Repair Orders a Month

A service department running 200 repair orders a month can reasonably manage manual detractor triage — if the flagging process is disciplined and the escalation path is defined. At 400, 600, or 1,000 ROs a month, the math breaks down. A BDC team placing 40–50 outbound calls per rep per day is already stretched to cover standard follow-up volume; the flagged-RO priority list competes directly with the regular queue and gets deprioritized when headcount is the binding constraint.

The contact rate ceiling is what actually limits the process, not the prioritization logic. Even a perfectly designed triage system fails if the BDC reaches 15–20% of customers in the first 24 hours — because the other 80–85% receive the survey email before anyone from the store has spoken to them. Flagging the high-risk ROs addresses the prioritization problem. It doesn't address the coverage problem. Both matter, but coverage determines whether a detractor's survey gets submitted before or after your call reaches them.

Cox Automotive's 2025 data shows that service retention among owners of two-year-or-newer vehicles dropped from 72% to 54% in just two years (Cox Automotive, 2025). Customers who do service at the dealership are 74% more likely to repurchase there — but 45% are currently dissatisfied with the service experience (Cox Automotive, 2025). The dealerships closing that gap aren't building more elaborate manual processes. They're solving the coverage problem: reaching 65–75% of customers inside the survey window instead of 15–20%, so the detractors who are recoverable actually get recovered.

“A perfectly designed triage system fails if the BDC only reaches 15–20% of customers in the first 24 hours. Flagging the high-risk ROs addresses prioritization. It doesn't address coverage — and coverage is what determines whether the survey gets submitted before or after your call.”

Frequently Asked Questions About Catching Detractors Before CSI Surveys

When exactly do OEM surveys get sent after a service visit? Industry practice, consistent across multiple sources, is that OEM surveys ship within 48–72 hours of the repair order closing. Most customers receive the survey 3–10 days after their service visit (J.D. Power, 2026). The exact timing is not published by OEMs directly and varies by brand — but the implication is consistent: your effective recovery window starts the moment the RO closes, not when the BDC gets around to the call list.

Can a recovery call actually change a customer's survey score? Yes — if it happens before the customer submits. TARP research shows that 54–70% of customers who lodge a complaint will return when the complaint is resolved, rising to 95% with same-day resolution (TARP). In our experience, the customers most likely to update their view are those with a specific, fixable grievance who were contacted by a manager with real authority. A recovery call after submission has no effect on that score — but it still matters for the long-term relationship.

Should my BDC rep ask the customer to hold off on submitting the survey? No. Asking a customer to delay submission or rate the experience more favorably is a violation of OEM survey integrity policies and creates a credibility risk if the customer perceives it as pressure. The goal of the recovery call is to resolve the problem, not to manage the form. If the problem is genuinely resolved and the customer feels heard, the survey outcome typically follows.

Which repair orders should I flag as highest priority? Flag four categories before building the day's call list: first-visit fix failures (returned repairs or unresolved complaints noted at pickup), excessive wait time (jobs more than 50% over quoted time), advisor escalations logged in the DMS, and repeat warranty work on the same issue. These go to the front of the queue regardless of alphabetical or chronological sort order — they have the shortest effective window before the survey ships.

What's the difference between a BDC follow-up call and a service recovery call? A BDC follow-up call confirms satisfaction and closes the loop on the visit. A service recovery call happens when a complaint signal is detected, and its purpose is to resolve the specific issue — with a manager on the line who has the authority to make things right. The wrong person making the recovery call — an advisor or BDC rep without resolution authority — typically makes the situation worse, not better. The handoff to a manager needs to happen within the same call or within two hours of the complaint being surfaced.